Classification of multiple cancer types by multicategory 
support vector machines using gene expression data

Lee, Yoonkyung; Lee, Cheol Koo

doi:10.1093/bioinformatics/btg102

Cited by 308 publications

(133 citation statements)

References 27 publications

Supporting

Mentioning

126

Contrasting

Unclassified

Order By: Relevance

“…The quality of the resulting dimensionally reduced data sets were tested using a support vector machine classifier, chosen because of its known suitability to this task [22][23][24]. The data mining package Weka [21] contains all algorithms used, apart from TPP which is made available in an extended version of the package on the associated website, along with all data used.…”

Section: Methodsmentioning

confidence: 99%

Feature Selection with Targeted Projection Pursuit

Enshaei

Faith²

2015

IJITCS

View full text Add to dashboard Cite

Abstract-The selection of attributes becomes more important, but also more difficult, as the size and dimensionality of data sets grows, particularly in bioinformatics. Targeted Projection Pursuit is a dimension reduction technique previously applied to visualising high-dimensional data; here it is applied to the problem of feature selection. The technique avoids searching the powerset of possible feature combinations by using perceptron learning and attraction-repulsion algorithms to find projections that separate classes in the data. The technique is tested on a range of gene expression data sets. It is found that the classification generalisation performance of the features selected by TPP compares well with standard wrapper and filter approaches, the selection of features generalises more robustly than either, and its time efficiency scales to larger numbers of attributes better than standard searches.

show abstract

Section: Methodsmentioning

confidence: 99%

Feature Selection with Targeted Projection Pursuit

Enshaei

Faith²

2015

IJITCS

View full text Add to dashboard Cite

show abstract

“…The research result shows the feasibility of establishing the models with FTIR-CWT-SVM method to identify normal, early carcinoma and advanced gastric carcinoma. (Mingjun and Rajasekaran, 2010) gives a greedy algorithm for gene selection (Lee and Lee, 2003) based on SVM and correlation. Microarrays serve scientists as a powerful and efficient tool to observe thousands of genes and analyze their activeness in normal or cancerous tissues.…”

Section: Introductionmentioning

confidence: 99%

Cancer Classification using Support Vector Machines and Relevance Vector Machine based on Analysis of Variance Features

Bharathi¹

2011

Journal of Computer Science

View full text Add to dashboard Cite

Problem statement:The objective of this study is, to find the smallest set of genes that can ensure highly accurate classification of cancer from micro array data by using supervised machine learning algorithms. The significance of finding the minimum subset is three fold: The computational burden and noise arising from irrelevant genes are much reduced; the cost for cancer testing is reduced significantly as it simplifies the gene expression tests to include only a very small number of genes rather than thousands of genes; it calls for more investigation into the probable biological relationship between these small numbers of genes and cancer development and treatment. Approach: The proposed method involves two steps. In the first step, some important genes were chosen with the help of Analysis of Variance (ANOVA) ranking scheme. In the second step, the classification capability was tested for all simple combinations of those important genes using a better classifier. Results: The proposed method initially uses Support Vector Machine (SVM) classifier. Relevance Vector Machine (RVM) classifier was used for increasing the classification accuracy over SVM classifier. Conclusion:The experimental result shows that the proposed method performs the cancer classification with better accuracy when compared to the conventional methods.

show abstract

“…These approaches range from traditional methods, such as Fisher's linear and quadratic discriminant analysis, to more modern machine learning techniques, such as classification trees or aggregation of classifiers by bagging or boosting (for a review see [14]) [12]. There are also approaches which are able to identify test samples that do not belong to any of the known classes by imposing thresholds on the prediction strength [13,15].…”

Section: Introductionmentioning

confidence: 99%